HIV Resistance: Data and spin.

On December 18 the first report was presented from a new study of the prevalence of drug-resistant HIV in U.S. patients in early 1999. (1) This study found that somewhere between 50 and 78 percent of these patients (depending on how you count patients whose viral resistance could not be measured) had some degree of reduced susceptibility to at least one antiretroviral. White, gay, middle class, insured patients had the most resistance, on the average, while those with less access to care had less. The national press eagerly picked up that story; and when we got home from the ICAAC conference in Chicago where the preliminary report was presented, we found that people all over the country had heard it -- and little else from the conference.

A closer look shows that while the study results are valid (though not as surprising as they might appear), the central messages that carried the press story appear to be misinterpretations -- ones that could have future consequences for society's political will to deal with the HIV epidemic, both in the U.S. and abroad:

1. The main message that went out through the press is that drugs are not working because of resistance. In fact, as one of the researchers noted to AIDS Treatment News, the good news is that treatments are still saving lives despite viral drug resistance. And most of the press ignored the fact, brought out at a press conference at ICAAC, that many of the patients found to have resistant virus started antiretrovirals years ago with inadequate regimens, and added new drugs one at a time as they became available in the 1990s -- conditions that facilitate resistance development. Patients starting treatment today do not use drugs that way.

2. The publicly available abstract of the study, as well as statements to the press, correctly reported that resistance was associated with markers of access to care. (Those with good access to medical care usually started treatment earlier, and therefore had more time to develop resistance -- and also they often started with the suboptimal two-drug or one-drug regimens.) But the emotional subtext that sold the newspapers was the implication that gay white men, despite all their advantages, were not doing their part to control the epidemic.

How the Study Was Done

This resistance study used samples collected in a major national survey of HIV care in the U.S., the HCSUS study (HIV Cost and Services Utilization Study). (2) The importance of HCSUS is that while most studies describe the particular patients who are available for the researchers (through a particular medical institution or clinical trial, for example), HCSUS carefully selected a sample to be as representative as possible of all HIV-positive persons receiving medical care in the U.S. (except in the military, in prison, or in a hospital emergency department), in the first two months of 1996. HCSUS randomly selected 4042 patients and interviewed 76% of them. It found that in January and February of 1996, about 230,000 HIV-infected adults received medical care. (2) HCSUS also found that "the patient population was disproportionately male, black, and poor," that many Americans with HIV were receiving care less than twice a year, and that the total cost of medical care for Americans with HIV was less than 1% of all direct personal health expenditures. (2)

In the new resistance study, over 1900 plasma samples obtained from HCSUS volunteers about three years later (in late 1998 to early 1999) were analyzed using the ViroLogic PhenoSense resistance test. Sixty-three percent of these samples had a viral load of over 500 copies of HIV, and 89% of those had resistance test results (those with a viral load lower than 500 cannot be tested for resistance with standard tests). Of those who could be tested, 78% had reduced susceptibility to at least one antiretroviral.

There was confusion in news reports over whether resistance was found in 78% of the patients, or in about half of them. This is because the most conservative calculation assumed no resistance in any of the patients who could not be tested for resistance. Therefore, 78% (of those successfully tested who were found to be resistant to at least one antiretroviral) times 63% (of those eligible for resistance testing since they had a viral load of over 500 copies), gives 49% of the total study population in which reduced susceptibility to at least one antiretroviral was documented. (This calculation is approximate, because in the actual study weighting factors were used to make the sample of patients studied be more representative of the U.S. HIV patient population.) Those who could not be tested probably tended to have less resistance than the others (since most had a low viral load, indicating the drugs were probably working well), but certainly persons with viral load under 500 can have drug-resistant virus.

This study did not collect adherence information except for self-reports, and does not have enough data to look at adherence.

AIDS Treatment News talked with Dr. Nick Hell-mann of Virologic, one of the authors of the resistance report. He noted that despite this viral resistance, the death rate in the U.S. has still been kept relatively low since modern combination treatment was introduced. He suspects that part of the reason is that unlike bacteria, HIV usually pays a significant price for drug resistance, and is likely to become less able to replicate and cause rapid worsening of disease. He noted that while it might be possible for HIV to evolve to be both highly resistant and highly pathogenic, this appears to be uncommon.

Comment

This study did indeed find more resistance (HIV with reduced susceptibility to antiretrovirals) than expected. But much of this result is not really surprising given the study design. The patients selected were all in care in the U.S. in early 1996, but had their blood drawn and virus tested three years later in late 1998 to early 1999. With this sampling, many of the patients would have been on antiretrovirals for a long time, giving more time for resistance to occur. Since all were in care in early 1996 and known at that time to have HIV, it is likely that many of them started on suboptimal therapies. This selection (plus the fact that resistance was tested for many drugs, and just one positive test led to the volunteer being counted as having resistant virus) may partly explain why this study found much more resistance than other studies.

The groups that started treatment earlier -- including gay men, and those with insurance -- had more resistance, probably because they had more time for it to develop (as well as more chance of having been exposed to the two-drug or one-drug antiretroviral regimens no longer in use).

Could the new publicity on high prevalence of resistance contribute to the arguments against providing antiretroviral treatment in Africa? This study only looked at the U.S. But it is reasonable to assume that if treatment is introduced correctly in African countries, the results of this U.S. study would not apply. There will be less resistance than in the U.S., if patients are started on modern regimens and managed correctly.

Also, the kinds of HIV that are not native to the U.S. (but have been common for years in Africa and other parts of the world) have not spread here to any large extent. Quite likely the major reason is those at risk of HIV in the U.S. are far more likely to get infected by a native virus, which probably blocks infection by other EN strains. So the media image of resistant "superviruses" spreading from Africa throughout the world is contrary to the facts observed for years.

The right message to take from this study is that viral resistance is a serious problem, and people should be more careful to use antiretrovirals correctly. It is also important to prevent transmission of resistant virus to persons who are HIV-negative. For those already infected, generally it is best to have HIV fully suppressed whenever antiretrovirals are used, so that there is little or no viral replication, and resistant virus cannot evolve. But for many patients this goal is not feasible. For these patients and for everyone else with HIV, we need new drugs that are more effective, less toxic, and less susceptible to viral resistance. We especially need new classes of treatments, including new targets for antiretrovirals, and immune-based therapies to help the body itself control HIV.

References

(1.) Richman DD, Bozzette S, Morton S, Chien S, Wrin T, Dawson K, and Hellmann N. The prevalence of antiretroviral drug resistance in the U.S. 41st International Conference on Antimicrobial Agents and Chemotherapy, Chicago, December 18 [abstract LB-17].

(2.) Bozzettee SA, Berry SH, Duan N, Richman D and others. The care of HIV-infected adults in the United States. The New England Journal of Medicine. December 24, 1998; volume 339, number 26, pages 1897-1904.